Dealing with Spatial Data using R

Alessia Calafiore

Overview

Introducing R and Spatial Data
Mapping Gender Inequality in Scotland

R and Spatial Data

Simple Features for R (the sf package)

From Edzer Pebesma vignette:

sf implements a widely used standard in spatial databases that describes how objects in the real world can be represented in computers;

A feature can be thought of as a thing, or an object in the real world, such as a building or a tree.

Features have a geometry describing where on Earth the feature is located, and they have attributes, which describe other properties.

The most common simple features for spatial analysis are POINTS, LINESTRING, POLYGONS

Once installed sf can be loaded as any other R library

library(sf)

Now let’s load some spatial data

boundaries <- st_read("scottish_wards.gpkg")

Reading layer `all_scotland_wards_4th' from data source 
  `C:\Users\Alessia\Documents\rLadiesSpatial\scottish_wards.gpkg' 
  using driver `GPKG'
Simple feature collection with 353 features and 6 fields
Geometry type: MULTIPOLYGON
Dimension:     XY
Bounding box:  xmin: 5513 ymin: 530249.9 xmax: 470327.5 ymax: 1220301
Projected CRS: OSGB36 / British National Grid

class(boundaries)

[1] "sf"         "data.frame"

Coordinate Reference Systems

Coordinates correspond to specific location on the Earth’s surface on the basis of a coordinate reference system (CRS). A common way to identify a CRS is through its EPSG code.

st_crs(boundaries)[[1]] #EPSG = 27700

[1] "OSGB36 / British National Grid"

CRS can be transformed with sf.

boundaries_WGS84 <- st_transform(boundaries, 4326)
st_crs(boundaries_WGS84)[[1]]

[1] "EPSG:4326"

Geometry operations (some)

On a single object

st_centroids()
st_buffer()
st_simplify()
st_union()

Between two objects

st_intersection()
st_difference()
st_union()

Spatial Join

Spatial joins are useful when you want to join attributes of one sf obj with an other object based on some geometry operation.

pts <- st_sample(boundaries[1:2, ], 10) # creates sample points within boundaries
pts <- st_as_sf(pts) #transform it into an sf object 
names(pts)

[1] "x"

names(boundaries[1:2, ])

[1] "CODE"     "Ward_No"  "Review"   "Council"  "ONS_2010" "Name"     "geom"

As you can see there is no attribute we could use to join these two datasets.

but can we do it spatially?

Let’s try then

pts_data<-st_join(pts, boundaries[1:2,])
head(pts_data)

Simple feature collection with 6 features and 6 fields
Geometry type: POINT
Dimension:     XY
Bounding box:  xmin: 384919.8 ymin: 805037 xmax: 394237.3 ymax: 813326.5
Projected CRS: OSGB36 / British National Grid
    CODE Ward_No     Review       Council  ONS_2010
1 00QAMA       1 4th review Aberdeen City S13002476
2 00QAMA       1 4th review Aberdeen City S13002476
3 00QAMA       1 4th review Aberdeen City S13002476
4 00QAMA       1 4th review Aberdeen City S13002476
5 00QAMM      12 4th review Aberdeen City S13002487
6 00QAMM      12 4th review Aberdeen City S13002487
                          Name                  geometry
1 Dyce / Bucksburn / Danestone POINT (386012.5 809104.4)
2 Dyce / Bucksburn / Danestone POINT (384919.8 813326.5)
3 Dyce / Bucksburn / Danestone     POINT (386888 811398)
4 Dyce / Bucksburn / Danestone POINT (386363.9 810260.3)
5            Torry / Ferryhill   POINT (394096.6 805037)
6            Torry / Ferryhill POINT (394237.3 805274.5)

Mapping Gender Inequality in Scotland

Gender Pay Gap and Gender Inequality

Broadly, the Gender Pay Gap is the difference in pay between women and men.
Globally, women on average, are paid about 20 per cent less than men (ILO).
Gender inequality in job opportunities and career.

Measuring Gender Inequality in Scotland

Socio-economic Classification (SeC - Scottish Census 2011)

C1 - Higher managerial, administrative and professional occupations

C2 - Lower managerial, administrative and professional occupations

C3 - Intermediate occupations

C4 - Small employers and own account workers

C5 - Lower supervisory and technical occupations

C6 - Semi-routine occupations

C7 - Routine occupations

C8 - Never worked and long-term unemployed

Let’s load some data

data <- read.csv("LC6121SC.csv")
names(data)

 [1] "Name"   "Gender" "AllPpl" "C1"     "C2"     "C3"     "C4"     "C5"    
 [9] "C6"     "C7"     "C8"

With these data let’s create a simple equality score (E) to compare the likelihood of belonging to each of these SeC between females and males.

A bit of data wrangling

library(tidyverse)

ward_data <- data %>%
  #filter out country level data
  filter(Name != "Scotland") %>%
  #reshape the data
  pivot_wider(id_cols = Name,
              names_from = Gender,
              values_from = starts_with("C")) %>%
  #compute the score
  mutate(
    C1_gap = ((C1_Females / C1_All) / (C1_Males / C1_All)) * 100,
    C2_gap = ((C2_Females / C2_All) / (C2_Males / C2_All)) * 100,
    C3_gap = ((C3_Females / C3_All) / (C3_Males / C3_All)) * 100,
    C4_gap = ((C4_Females / C4_All) / (C4_Males / C4_All)) * 100,
    C5_gap = ((C5_Females / C5_All) / (C5_Males / C5_All)) * 100,
    C6_gap = ((C6_Females / C6_All) / (C6_Males / C6_All)) * 100,
    C7_gap = ((C7_Females / C7_All) / (C7_Males / C7_All)) * 100,
    C8_gap = ((C8_Females / C8_All) / (C8_Males / C8_All)) * 100
) %>%
select(Name, contains("gap"))

How to interpret the score (E)

If E = 100 females and males have the same likelihood to be in a SeC;
If E > 100 it is more likely for females than males to be in a SeC;
if E < 100 it is more likely for males than females to be in a SeC;

Let’s get some descriptive stats

ward_data %>%
  pivot_longer(cols = contains("gap"),
               names_to = "SeC",
               values_to = "Gap") %>%
  group_by(SeC) %>%
  summarise(
    Min = min(Gap),
    Max = max(Gap),
    Mean = mean(Gap),
    SD = sd(Gap)
  )

# A tibble: 8 × 5
  SeC      Min   Max  Mean    SD
  <chr>  <dbl> <dbl> <dbl> <dbl>
1 C1_gap  37.0  84.9  55.1  9.22
2 C2_gap  81.3 177.  134.  14.2 
3 C3_gap  66.8 510.  287.  57.4 
4 C4_gap  22.2  74.0  43.8  9.55
5 C5_gap  23.8  64.9  38.2  5.90
6 C6_gap  96.0 292.  182.  27.6 
7 C7_gap  34.6 113.   68.2 10.9 
8 C8_gap  45.4 233.  110.  23.9

Static maps with ggplot2

ggplot is a widely used R library, now over 10 years old

It is used to make any type of plot, including maps !!!

Browse the R gallery to find out more

Let’s make some maps !!!

First we need to link the boundaries with the data

names(boundaries)

[1] "CODE"     "Ward_No"  "Review"   "Council"  "ONS_2010" "Name"     "geom"

names(ward_data)

[1] "Name"   "C1_gap" "C2_gap" "C3_gap" "C4_gap" "C5_gap" "C6_gap" "C7_gap"
[9] "C8_gap"

Let’s join these two datasets

ward_data_geo <- boundaries %>%
  inner_join(ward_data, by = c("Name"="Name"))
names(ward_data_geo)

 [1] "CODE"     "Ward_No"  "Review"   "Council"  "ONS_2010" "Name"    
 [7] "C1_gap"   "C2_gap"   "C3_gap"   "C4_gap"   "C5_gap"   "C6_gap"  
[13] "C7_gap"   "C8_gap"   "geom"

Now we can make our first map

ggplot()+
  geom_sf(aes(fill = C1_gap),data = ward_data_geo)

Let’s make some changes

ggplot()+
  geom_sf(aes(fill = C1_gap),
          data = ward_data_geo,
          colour = "NA") +
  #change color palette
  scale_fill_gradientn( 
            colors = c("#0002A1", "#332FD0", "#FB2576", "#3F0071")
                ) +
  #add a scale bar
  ggspatial::annotation_scale( 
    location = "bl",
    bar_cols = c("grey60", "white")
    ) +
  #change text 
  labs(fill = "Women every 100 Men\nin Higher Managerial Roles",
       title = "Gender Inequality in Scotland")+
  #customize theme
  theme_void() + 
  theme(
    text = element_text(family = "Futura-Medium",color = "#22211d"),
    legend.title = element_text(family = "Futura-Bold", size = 10),
    legend.text = element_text(family = "Futura-Medium", size = 10),
    plot.background = element_rect(fill = "#f5f5f2", color = NA),
    panel.background = element_rect(fill = "#f5f5f2", color = NA)
    )

ggplot()+
  geom_sf(aes(fill = C1_gap),
          data = ward_data_geo,
          colour = "NA") +
  #change color palette
  scale_fill_gradientn( 
            colors = c("#0002A1", "#332FD0", "#FB2576", "#3F0071")
                ) +
  #add a scale bar
  ggspatial::annotation_scale(  
    location = "bl",
    bar_cols = c("grey60", "white")
    ) +
  #change text 
  labs(fill = "Women every 100 Men\nin Higher Managerial Roles",
       title = "Gender Inequality in Scotland")+
  #customize theme
  theme_void() + 
  theme(
    text = element_text(family = "Futura-Medium",color = "#22211d"),
    legend.title = element_text(family = "Futura-Bold", size = 10),
    legend.text = element_text(family = "Futura-Medium", size = 10),
    plot.background = element_rect(fill = "#f5f5f2", color = NA),
    panel.background = element_rect(fill = "#f5f5f2", color = NA)
    )

ggplot()+
  geom_sf(aes(fill = C1_gap),
          data = ward_data_geo,
          colour = "NA") +
  #change color palette
  scale_fill_gradientn( 
            colors = c("#0002A1", "#332FD0", "#FB2576", "#3F0071")
                ) +
  #add a scale bar
  ggspatial::annotation_scale(  
    location = "bl",
    bar_cols = c("grey60", "white")
    ) +
  #change text 
  labs(fill = "Women every 100 Men\nin Higher Managerial Roles",
       title = "Gender Inequality in Scotland")+
  #customize theme
  theme_void() + 
  theme(
    text = element_text(family = "Futura-Medium",color = "#22211d"),
    legend.title = element_text(family = "Futura-Bold", size = 10),
    legend.text = element_text(family = "Futura-Medium", size = 10),
    plot.background = element_rect(fill = "#f5f5f2", color = NA),
    panel.background = element_rect(fill = "#f5f5f2", color = NA)
    )

and see the result

Focus on Edinburgh

Preparing data to add ward labels

#create polygon centroids to be used as location for the labels
label_poi <- ward_data_geo %>%
  filter(Council %in% c("City of Edinburgh")) %>%
  st_centroid()

#add long lat field
label_poi <- label_poi %>%
  mutate(LONG=as.numeric(st_coordinates(label_poi)[,1]),
         LAT=as.numeric(st_coordinates(label_poi)[,2]))

names(label_poi)

 [1] "CODE"     "Ward_No"  "Review"   "Council"  "ONS_2010" "Name"    
 [7] "C1_gap"   "C2_gap"   "C3_gap"   "C4_gap"   "C5_gap"   "C6_gap"  
[13] "C7_gap"   "C8_gap"   "geom"     "LONG"     "LAT"

Let’s add the labels to the map

#filter the data
ward_data_geo %>%
  filter(Council %in% c("City of Edinburgh")) %>%
  ggplot() +
  geom_sf(aes(fill = C1_gap),
          colour = "#f5f5f2") +
  scale_fill_gradientn(
            colors = c("#0002A1", "#332FD0", "#FB2576", "#3F0071")
                ) +
  theme_void() +
  ggspatial::annotation_scale(
    location = "br",
    bar_cols = c("grey60", "white")
    ) +
  labs(fill = "Women for every 100 Men\nin Higher Managerial Roles",
       title = "Gender Inequality in Edinburgh")+
  #customize theme
  theme(
    text = element_text(family = "Futura-Medium",color = "#22211d"),
    legend.title = element_text(family = "Futura-Bold", size = 10),
    legend.text = element_text(family = "Futura-Medium", size = 10),
    plot.background = element_rect(fill = "#f5f5f2", color = NA),
    panel.background = element_rect(fill = "#f5f5f2", color = NA)
    )  +
  #add labels
  geom_sf(data = label_poi, color="NA")+
  ggrepel::geom_text_repel(mapping = aes(x=LONG,
                                         y=LAT,
                                         label=Name), 
                           data = label_poi,
                           size=2,
                           color = "white",     # text color
                           bg.color = "grey30", # shadow color
                           bg.r = 0.15,
                           max.overlaps=Inf,
                           segment.color = "black")

#filter the data
ward_data_geo %>%
  filter(Council %in% c("City of Edinburgh")) %>%
  ggplot() +
  geom_sf(aes(fill = C1_gap),
          colour = "#f5f5f2") +
  scale_fill_gradientn(
            colors = c("#0002A1", "#332FD0", "#FB2576", "#3F0071")
                ) +
  theme_void() +
  ggspatial::annotation_scale(
    location = "br",
    bar_cols = c("grey60", "white")
    ) +
  labs(fill = "Women for every 100 Men\nin Higher Managerial Roles",
       title = "Gender Inequality in Edinburgh")+
  #customize theme
  theme(
    text = element_text(family = "Futura-Medium",color = "#22211d"),
    legend.title = element_text(family = "Futura-Bold", size = 10),
    legend.text = element_text(family = "Futura-Medium", size = 10),
    plot.background = element_rect(fill = "#f5f5f2", color = NA),
    panel.background = element_rect(fill = "#f5f5f2", color = NA)
    )  +
  #add labels
  geom_sf(data = label_poi, color="NA")+
  ggrepel::geom_text_repel(mapping = aes(x=LONG,
                                         y=LAT,
                                         label=Name), 
                           data = label_poi,
                           size=2,
                           color = "white",     # text color
                           bg.color = "grey30", # shadow color
                           bg.r = 0.15,
                           max.overlaps=Inf,
                           segment.color = "black")

and see the result

Let’s add some interactivity

Interactive maps with leaflet

Leaflet is JavaScript library for mobile-friendly interactive maps.

This R package makes it easy to integrate and control Leaflet maps in R.

After installed you can load it:

library(leaflet)

Some data prep

#transform CRS to make it compatible with global map tiles prividers
ward_data_geo <- st_transform(ward_data_geo, 4326) %>%
  st_make_valid() %>%
  st_simplify()

#find centroid of Scotland to center the first view
scotland_ct <- ward_data_geo %>%
  st_make_valid() %>%
  st_union() %>%
  st_centroid()

Basic example

leaflet() %>%
  setView(scotland_ct[[1]][1], scotland_ct[[1]][2], zoom = 6) %>%
  # add a dark basemap
  addProviderTiles("CartoDB.DarkMatter") %>%
  # add the polygons of the clusters
  addPolygons(
    data = ward_data_geo,
    color = "#E2E2E2",
    # set the opacity of the outline
    opacity = 1,
    # set the stroke width in pixels
    weight = 1,
    # set the fill opacity
    fillOpacity = 0.2
  )

leaflet() %>%
  setView(scotland_ct[[1]][1], scotland_ct[[1]][2], zoom = 6) %>%
  # add a dark basemap
  addProviderTiles("CartoDB.DarkMatter") %>%
  # add the polygons of the clusters
  addPolygons(
    data = ward_data_geo,
    color = "#E2E2E2",
    # set the opacity of the outline
    opacity = 1,
    # set the stroke width in pixels
    weight = 1,
    # set the fill opacity
    fillOpacity = 0.2
  )

leaflet() %>%
  setView(scotland_ct[[1]][1], scotland_ct[[1]][2], zoom = 6) %>%
  # add a dark basemap
  addProviderTiles("CartoDB.DarkMatter") %>%
  # add the polygons of the clusters
  addPolygons(
    data = ward_data_geo,
    color = "#E2E2E2",
    # set the opacity of the outline
    opacity = 1,
    # set the stroke width in pixels
    weight = 1,
    # set the fill opacity
    fillOpacity = 0.2
  )

Let’s see what we get

Add fill and legend

bins <- c(35, 45, 55, 65, 75, 85)
pal <- colorBin(c("#0002A1", "#332FD0", "#FB2576", "#3F0071"), domain = ward_data_geo$C1_gap, bins = bins)
leaflet(ward_data_geo) %>%
  # center the map view
  setView(scotland_ct[[1]][1], scotland_ct[[1]][2], zoom = 6) %>%
  # add a dark basemap
  addProviderTiles("CartoDB.DarkMatter") %>%
  # change polygon fill
  addPolygons(
    # set the opacity of the outline
    opacity = 0,
    # set the stroke width in pixels
    weight = 0.1,
    # set fill colors
    fillColor = ~pal(C1_gap),
    # set the fill opacity
    fillOpacity = 0.8
  ) %>%
  addLegend(pal = pal, 
            values = ~C1_gap, 
            opacity = 0.7,
            position = "bottomright",
            title = "Women every 100 Men</br>in Higher Managerial Roles")

bins <- c(35, 45, 55, 65, 75, 85)
pal <- colorBin(c("#0002A1", "#332FD0", "#FB2576", "#3F0071"), domain = ward_data_geo$C1_gap, bins = bins)
leaflet(ward_data_geo) %>%
  # center the map view
  setView(scotland_ct[[1]][1], scotland_ct[[1]][2], zoom = 6) %>%
  # add a dark basemap
  addProviderTiles("CartoDB.DarkMatter") %>%
  # change polygon fill
  addPolygons(
    # set the opacity of the outline
    opacity = 0,
    # set the stroke width in pixels
    weight = 0.1,
    # set fill colors
    fillColor = ~pal(C1_gap),
    # set the fill opacity
    fillOpacity = 0.8
  ) %>%
  addLegend(pal = pal, 
            values = ~C1_gap, 
            opacity = 0.7,
            position = "bottomright",
            title = "Women every 100 Men</br>in Higher Managerial Roles")

bins <- c(35, 45, 55, 65, 75, 85)
pal <- colorBin(c("#0002A1", "#332FD0", "#FB2576", "#3F0071"), domain = ward_data_geo$C1_gap, bins = bins)
leaflet(ward_data_geo) %>%
  # center the map view
  setView(scotland_ct[[1]][1], scotland_ct[[1]][2], zoom = 6) %>%
  # add a dark basemap
  addProviderTiles("CartoDB.DarkMatter") %>%
  # change polygon fill
  addPolygons(
    # set the opacity of the outline
    opacity = 0,
    # set the stroke width in pixels
    weight = 0.1,
    # set fill colors
    fillColor = ~pal(C1_gap),
    # set the fill opacity
    fillOpacity = 0.8
  ) %>%
  addLegend(pal = pal, 
            values = ~C1_gap, 
            opacity = 0.7,
            position = "bottomright",
            title = "Women every 100 Men</br>in Higher Managerial Roles")

bins <- c(35, 45, 55, 65, 75, 85)
pal <- colorBin(c("#0002A1", "#332FD0", "#FB2576", "#3F0071"), domain = ward_data_geo$C1_gap, bins = bins)
leaflet(ward_data_geo) %>%
  # center the map view
  setView(scotland_ct[[1]][1], scotland_ct[[1]][2], zoom = 6) %>%
  # add a dark basemap
  addProviderTiles("CartoDB.DarkMatter") %>%
  # change polygon fill
  addPolygons(
    # set the opacity of the outline
    opacity = 0,
    # set the stroke width in pixels
    weight = 0.1,
    # set fill colors
    fillColor = ~pal(C1_gap),
    # set the fill opacity
    fillOpacity = 0.8
  ) %>%
  addLegend(pal = pal, 
            values = ~C1_gap, 
            opacity = 0.7,
            position = "bottomright",
            title = "Women every 100 Men</br>in Higher Managerial Roles")

Add popup

bins <- c(35, 45, 55, 65, 75, 85)
pal <- colorBin(c("#0002A1", "#332FD0", "#FB2576", "#3F0071"), domain = ward_data_geo$C1_gap, bins = bins)
leaflet(ward_data_geo) %>%
  # center the map view
  setView(scotland_ct[[1]][1], scotland_ct[[1]][2], zoom = 6) %>%
  # add a dark basemap
  addProviderTiles("CartoDB.DarkMatter") %>%
  # change polygon fill
  addPolygons(
    # set the opacity of the outline
    opacity = 0,
    # set the stroke width in pixels
    weight = 0.1,
    # set fill colors
    fillColor = ~pal(C1_gap),
    # set the fill opacity
    fillOpacity = 0.8,
    # add popup
    popup = paste('<strong>',"Gap: ",'</strong>', round(ward_data_geo$C1_gap), "<br>",
                  '<strong>',"Ward Name:",'<strong>', ward_data_geo$Name, "<br>",
                  '<strong>',"Council: ",'<strong>', ward_data_geo$Council, "<br>")
  ) %>%
  addLegend(pal = pal, 
            values = ~C1_gap, 
            opacity = 0.7,
            position = "bottomright",
            title = "Women every 100 Men</br>in Higher Managerial Roles")

Conclusions

These are not research results, but just some quick data exploration to showcase the R magic !

More affluent and urban areas seem to have lower differences between women and men.

It would be interesting to explore this more as it demonstrate the importance of intersectional feminism perspectives.

There are many more ways/libraries that can be used to make maps.

Hope you enjoyed the tutorial

As always

Contacts

acalafio@ed.ac.uk

@alel_domi

@alel@datasci.social